Effect of varying Alternanthera philoxeroides ( alligator weed ) cover on the macrophyte species diversity of pond ecosystems : a quadrat – based study

This year-long study, covering three main seasons of India, focused on enumerating the effect of varying cover of Alternanthera philoxeroides (alligator weed) on the associated macrophyte species diversity of the littoral region of natural pond ecosystem. A total of 192 quadrats were randomly placed in the littoral region of 12 similar ponds containing varying degrees of A. philoxeroides infestation to estimate A. philoxeroides ‘cover (%)’ and number of associated macrophyte species in each quadrat. Overall, 20 associated macrophyte species, including 16 aquatic/ littoral-associated, 2 non-aquatic species, and grasses and sedges, were found to be present. A. philoxeroides infestation was categorized into 4 cover grades (Grade I-IV) from lowest (no/negligible: <10% cover) to highest (>60% cover). For each season, significant differences in the total number of associated macrophyte species across the 4 A. philoxeroides cover grades were found. A Poisson-regression model showed that for each season, even when the effect of other invasive species was adjusted, the number of associated native macrophyte species in a quadrat decreases significantly with increase in A. philoxeroides cover. A comparison of the quadrat communities between the lowest grade (Grade I) and highest grade (Grade IV) of A. philoxeroides infestation showed a significant reduction of species richness, diversity and evenness from the lowest to the highest infestation grades. Again, Mann-Whitney U tests further revealed that the number of native macrophyte species was significantly lower at highest (Grade IV) A. philoxeroides infestation than that at lowest infestation (Grade I). The presence of multiple invaders in A. philoxeroides infested aquatic ecosystems is also reported, indicating probable facilitative interactions between A. philoxeroides and other invasive species. The socio-economic valuation of some important native plants, which were found to be significantly reduced at high infestation levels of A. philoxeroides, has also been highlighted.


Introduction
The introduction and spread of non-native species has become a global ecological and conservation crisis as invasive organisms are increasingly altering terrestrial and aquatic communities worldwide, causing catastrophes for the native ecosystems (McNeely 2001), leading to biodiversity loss (Powell et al. 2011) and local extinction of native species (Butchart et al. 2010).Invasive species may inflict harmful ecological and economic impacts upon ecosystems in non native regions (Pimentel et al. 2005).They can lead towards ecosystem degradation and impairment of ecosystem services worldwide (Pysek and Richardson 2010), thereby inducing high conservation concerns (Wilcove et al. 1998).Direct consequences of recent anthropogenic activities like environmental modifications (Vitousek et al. 1997) and worldwide increase in trade and transport (Lockwood 1999) are the leading causes of creation of invasive species on a global scale.
Aquatic macrophytes are key components of aquatic and wetland ecosystems (Rejmánková 2011) playing a pivotal role in the ecosystem, like oxygenation of water (Caraco et al. 2006), productivity and nutrient retention (Engelhardt and Ritchie 2001), providing shelter and refuge and food (Wetzel 2001) for aquatic macroinvertebrates and other animals.Native aquatic macrophytes not only form an important source of food and medicine but also provide livelihood for many marginalised people and are essential to society (Ghosh 2005).Although freshwater ecosystems cover approximately 0.01% of the earth's water, they contain about 6% of all described species (Dudgeon et al. 2006), have the highest extinction rates (Jenkins 2003) and are under extreme pressure (Michelan et al. 2010).The problem is compounded by aquatic invasive species, which are known to have dramatic effects where they become introduced and established (Homans and Smith 2013) causing huge changes in many ecosystems worldwide and profoundly altering native communities and ecosystems (Gurevitch and Padilla 2004).The full impact of invasive aquatic species can only be evaluated when its effect on human well-being, such as the number of people affected and the magnitude of its impact on their lives (Pejchar and Mooney 2004) are also accounted.Loss of native biodiversity not only threatens the productivity and sustainability of ecosystems (Tilman et al. 1996), but also has significant socioeconomic and environmental impacts, especially in developing, over-populated countries of Asia and Africa (Juffe-Bignoli et al. 2012).
Alternanthera philoxeroides (Martius) Griseb (Amaranthaceae), commonly known as 'alligator weed', is an invasive perennial wetland herb originating from the Parana River region of South America (Maddox 1968).Cross-continental transport primarily by ships' ballast water until the turn of the 20 th century (Gunasekera and Bonilla 2001) and deliberate and accidental inter-catchment dispersals by humans in present times (Sainty et al. 1998), are considered to be the primary pathways of its global spread.A. philoxeroides fulfils most criteria of Baker's (1974) ideal weed characteristics and is regarded as one of the worst weeds of the world because of its invasiveness, potential for spread, high tolerance to environmental fluctuations and a wide range of adaptive potential (Chatterjee and Dewanji 2012) and has detrimental economic and environmental impacts.Once established, it is extremely hard to control and eradication is very expensive (Sainty et al. 1998).In spite of the recognition of A. philoxeroides as a problematic invasive plant (Wang et al. 2008), field-based studies about the ecological consequences of this species have been relatively few. A. philoxeroides invasion have been associated with altered soil decomposition dynamics and composition of soil microorganisms (Bassett et al. 2011) and allelopathic inhibition of water blooms (Zuo et al. 2012).A majority of the studies report the presence of A. philoxeroides in different ecosystems (Chatterjee and Dewanji 2010;Masoodi and Khan 2012) and model its spread in different aquatic ecosystems like lakes (Erwin et al. 2013;Masoodi et al. 2013) and ponds (Clements et al. 2011).Assessments of the impacts of individual invasive species on the native aquatic ecosystems are also scarce (Oreska and Aldridge 2011) and much work remains to be done (Vilà et al. 2010) especially in the invaded ecosystems of the tropics and subtropics.
This study aimed to investigate the impact of A. philoxeroides invasion on the associated macrophyte diversity of littoral zones of natural pond ecosystems.Ideally the experimental design for such a case-control study should have included different pond ecosystems with and without A. philoxeroides invasion for measurement of the plant community richness.But most of the ponds in our area already contained A. philoxeroides.Thus, this study was restricted to an 'impact-only' analysis of A. philoxeroides infestation on the floristic composition of aquatic/littoral macrophytes.We hypothesised that there would be a negative impact of increasing A. philoxeroides abundance, measured as cover percentage, on the overall associated macrophyte diversity of any pond ecosystem, across the 3 main seasons of India.Specifically, the effect of no/negligible and very high A. philoxeroides infestation on the associated native macrophyte diversity was studied.The study also presents a detailed floristic account of some native aquatic/ littoral macrophytes of India along with their socio-ecological value, and tries to enumerate the effect of A. philoxeroides infestation on the growth of these plants, and in turn, highlight some socio-economic implications of such invasion of pond ecosystems in the developing world.

Study area and duration
The study was conducted in a highly populated suburban locality (Baranagar) in the northern part of Greater Kolkata Metropolitan area.After surveying a large number of ponds within a radius of 10 kms, 12 ponds, which were homogenous in terms of average water-quality (pH: 6.5 -8.5; Conductivity < 1000 µS/cm; Turbidity < 10 NTU; Total Phosphorus < 5 mg/L; Total Kjeldahl Nitrogen < 10 mg/L; Dissolved Oxygen > 5 mg/L) and other anthropogenic influences, were selected for the study with the primary ascertainment such that varying degrees of A. philoxeroides cover were represented in the selected sample of the ponds.Due to logistic problems in obtaining suitable ponds as controls as mentioned before, a case-control study could not be designed.Instead, the primary focus of this research was directed to an "impact only" study where the severity of varying degree of A. philoxeroides on associated macrophyte richness in the littoral zones were explored.
The study was conducted over a period of 1 year (2009)(2010), covering all the three major seasons of India, viz.: summer (March -June), monsoon (July -October) and winter (November -February).The mean seasonal average temperatures were 31.33 ± 1.81ºC, 29.94 ± 1.46ºC and 21.89 ± 3.63ºC respectively while the total rainfall varied considerably between seasons (398.2 mm in summer, 866.9 mm during the monsoon and 21.1 mm in winter).Meteorological data for the duration of the study was obtained from the Regional Meteorological Centre, Kolkata, Indian Meteorological Department, Ministry of Earth Sciences, Government of India.

Sampling protocol
The standard procedure of visual estimation by the quadrat method (Jaccard 1901) was employed to record cover percent of A. philoxeroides as well as to estimate the macrophyte biodiversity (alpha diversity) within and near the littoral region of pond ecosystems.Following standardization, the quadrat size was fixed at 1 sq.meter while the minimum number of quadrats was found to be 11 for each of the ponds.To increase efficiency of estimation, 16 1m 2 quadrats were randomly placed in littoral zones of each pond per season.In each quadrat, the percentage area covered by A. philoxeroides was recorded as its cover (%) and the associated macrophyte species present therein were noted in order to check for differences between A. philoxeroides cover and number of associated macrophyte species.Identification of all macrophytes up to the specieslevel were done (Prain 1903;Cook 1996), except for the families 'Poaceae' and 'Cyperaceae', which were considered as single units, referred to as 'Grasses' and 'Sedges' respectively.
In order to highlight the socio-economic utility value of the native species with respect to that of A. philoxeroides, a subjective valuation scoring of these plants were done.Four categories of usage value were defined, primarily based on literature reviews, and scored under food, fodder, medicine and others (eg.aquaculture, fish feed, ornamental etc.).Further under food value, staple vegetables were given a score of 2, while vegetable supplements were scored as 1.Similarly when plants were popularly used as traditional medicines, they were given a score of 2, while the less commonly used ones were given a score of 1. Plants with fodder and 'other' miscellaneous use were both scored as 1.Additionally, four local suburban markets of Kolkata were surveyed to check the market availability of the plants as per their socio-economic utility.When plants were available, subjective scoring was done under two heads, namely 'Market Supply' -where daily supply was scored as 1.0, supply at intervals of 3-4 days was scored as 0.5, while a supply of once a week was given a score of 0.25; and 'Abundance in shops' -where presence in most shops were scored 1 against 0.5, when available in only a few specialized shops.

Statistical analysis
The basic sampling unit for the purpose of all analysis was chosen to be the quadrat, which resulted in a larger number of observations (n = 16 quadrats × 12 ponds = 192 quadrats per season) compared to a pond-based study, providing the flexibility of more reliable inferences.Each quadrat was considered to be a representative area of the littoral zone of the aquatic ecosystem, and they were classified into 4 infestation grades on the basis of A. philoxeroides cover.Figure 1  (A-D) presents the four cover groups of A. philoxeroides namely, 'No/Negligible infestation' (Grade I: 0-10% cover); 'Low-Medium infestation' (Grade II: 10-30% cover); 'Medium-High infestation' (Grade III: 30-60% cover) and 'Extremely High infestation' (Grade IV: 60-100% cover).In order to evaluate the simultaneous effects of sampling seasons and study ponds on the number of associated macrophyte species, a twoway analysis with suitable modifications to the standard ANOVA was carried out, since the variable of interest (associated macrophyte species) does not follow a normal distribution.Thus, instead of using the usual F-distributions, the empirical p-values corresponding to ANOVA statistics based on permutations were computed to analyse the data.This statistical test verified the effect (if any) of the different study ponds on the number of associated macrophyte species, so that the quadrat-wise study could be justified.The species accumulation curves computed for each of the 3 seasons were similar and revealed the adequacy of using 192 quadrats per season; the results for monsoon season is presented in Figure 2.
The Kruskal-Wallis H test was used to analyse the differences in overall cover (%) of  A. philoxeroides between 3 seasons.Again, for each sampling season, the Kruskal-Wallis H test was used to analyse whether the number of associated macrophyte species differed significantly across the four cover (%) groups or infestation grades of A. philoxeroides.The Mann-Whitney U test was used to evaluate the pair-wise differences for number of associated macrophyte species between the lowest (Grade I) and highest (Grade IV) infestation grades of A. philoxeroides.
To evaluate the effect of A. philoxeroides cover on native macrophyte diversity, a modified Poisson-regression model was used, given by: Y|X,Z ~ Poisson (e α+βX+γZ ) where Y = the number of native associated macrophyte species in the quadrat; X = the actual density of Alternanthera philoxeroides as cover % in the quadrat; Z = the number of other invasive macrophyte species in the quadrat.
We incorporated in the model, the number of other invasive macrophyte species in the quadrat (Z) as it could be a potential confounder in the association between the actual density (cover %) of A. philoxeroides (X) and the number of native macrophyte species (Y).This statistical test corrected for the effect of 'other invasive species' and gave the results as the exclusive effect of A. philoxeroides cover on native macrophyte diversity in the quadrats.
For each season, the Simpson's Diversity Index (D) (Simpson 1949), based on the species presence/absence data, was used to estimate the species diversity for the lowest and highest infestation grades of A. philoxeroides, as given by the formula: where S = the Species Richness (i.e. total number of species in that group); n = the total number of times the i-th species is present combining all the quadrats of that infestation grade; N = the total number of individuals (i.e. total count of the number of times each species is present combining all the quadrats of that infestation grade).
Similarly, for estimation of the corresponding evenness of these 2 groups, an evenness measure (E 1/D ), was calculated and termed as 'Simpson's Evenness Index' (Kent 2012).This was obtained by dividing the reciprocal form of the Simpson's Diversity Index (1/D) by the number of species in the sample (S), as given by the formula: The Microsoft excel, IBM-SPSS (ver.16) and R-language (R Core Team (2013) http://www.Rproject.org/)was used for statistical evaluation of the data.

Macrophyte diversity along infestation grades of A. philoxeroides
ANOVA results of the simultaneous effects of sampling seasons and study ponds on the number of associated species revealed neither significant  The overall species richness (S), Simpson's diversity index (D) and evenness index (E 1/D ) between the lowest (Grade I) and highest (Grade IV) infestation grade of A. philoxeroides were calculated and presented in Table 3.For each season, species diversity (D) was higher at Grade I (winter: 0.245, summer: 0.181, monsoon: 0.203) than the diversity of Grade IV (winter: 0.438, summer: 0.440, monsoon: 0.645).The evenness index (E 1/D ) which ranges from 0 (all biomass in one species) to 1 (>1 species present in equal abundance), with lower values indicating larger differences in abundance between species (Mulder et al. 2004), showed that at highest grade of infestation, evenness was lower (winter: 0.208, summer: 0.189, monsoon: 0.141) when compared with the corresponding evenness at Grade I (winter: 0.240, summer: 0.307, monsoon: 0.259).The results also showed that maximum reduction of species richness was during the monsoon (42%), followed by winter (35%) and summer (33%) from Grade I to Grade IV.
For each season, the exclusive effect of A. philoxeroides infestation impact on the associated macrophyte species richness, after adjusting for the number of other invasive species in that quadrat, was checked using the Poisson-regression model.The results of these tests revealed that the number of native macrophyte species in a quadrat decreased significantly with the increase in A. philoxeroides cover (monsoon: p<0.001; winter: p=0.014; summer: p=0.009), thereby indicating the negative effect of A. philoxeroides density on associated native species in littoral regions of the aquatic ecosystems.Additionally, the Mann-Whitney U test also revealed a significant decrease in number of native associated macrophytes across the lowest (Grade I) and the highest (Grade IV) of A. philoxeroides infestation for each season: U = 481.0,z = -3.744,p < 0.001 (winter); U=1566.0,z = -2.305,p = 0.021 (summer); U=175.5, z=-7.453,p < 0.001 (monsoon).The mean ranks of maximum number of native macrophyte species for Grade I were 74.02, 76.80 and 67.58; while the ranks for Group IV were 37.29, 59.65 and 24.89 for the seasons winter, summer and monsoon respectively.The frequency of overall quadrat-wise occurrence (%) of 12 trueaquatic native associated macrophytes across the lowest (Grade I) and highest (Grade IV) infestation grades, represented in Figure 4, shows that the occurrence of all these native macrophytes were impacted at higher infestation (Grade IV) of A. philoxeroides.The presence of dominant cooccurring native species like Ipomoea sp.(from 42.3% to 34.0%), Commelina sp.(from 21.1% to 6.2%) and Colocasia sp.(from 12.7% to 10.3%) were considerably reduced, while 50% of these 12 native associated macrophytes (Ceratophyllum sp., Otelia sp., Nymphaea sp., Enhydra sp., Nymphoides sp. and Marselia sp.) were completely eliminated at Grade IV of A. philoxeroides infestation.
We identified the overall socio-economic importance of some aquatic/littoral native associated macrophytes found in this study, including a subjective scoring of their sociological valuation and market availability, and compared it with A. philoxeroides (Table 4).It was noted that 10 out of these 12 native plants were used as vegetables or food supplements (Colocasia sp., Enhydra sp., Ipomoea sp., Ludwigia sp., Marsilea sp., Nymphaea sp., Nymphoides sp. and Ottelia sp., Polygonum sp., Vallisneria sp.).All these plants, except for Vallisneria sp., were also used as traditional medicines.Ipomoea sp., which had a high sociological usage score, was also the most popular native aquatic plant, having daily market turnover and wide-spread availability, followed by Colocasia sp. and Nymphaea sp.Again while Nymphoides sp. had a high sociological usage score; it was less popular in the market when compared to plants like Enhydra sp., Marselia sp. and Ludwigia sp.Although A. philoxeroides had some sociological use value at par with some native plants (Ceratophyllum sp., Ottelia sp., Vallisneria sp.Commelina sp. and Polygonum sp.), but unlike these natives, all of which were marginally popular in the market, none of the markets surveyed during the study sold A. philoxeroides on a commercial basis, indicating the low popularity of A. philoxeroides in the sub-urban markets; and hence emphasizing the higher importance of these native plants over A. philoxeroides in the society.

Discussion
A. philoxeroides, although perennial in the tropics and semi-tropics, exhibited low covers during the winter months.The monsoon season, with ambient temperature nearing 30°C and high seasonal rainfall (>850.0mm), provided the most favourable environmental conditions for the luxuriant growth of A. philoxeroides, which was evident by the significantly higher overall cover (%) of A. philoxeroides during this season.Good rainfall and temperatures around 30°C have been reported to be conducive for optimum shoot emergence and growth (Sainty et al. 1998;Shen et al. 2005) as well as peak plant height and maximum biomass (Liu et al. 2004) of A. philoxeroides.
Overall 18 aquatic/littoral macrophytes, along with grasses and sedges, were found in association with A. philoxeroides in the quadrats, of which 3 were other invasive plant species.Gradual decrease in total number of associated species (species richness) was observed from the lowest to highest infestation grades of A. philoxeroides, for each season.When comparing between the community structures of the quadrats belonging to the lowest (no/negligible) and highest infestation grades of A. philoxeroides, for all 3 seasons, reduction of both species diversity and evenness were revealed at the higher infestation grades of A. philoxeroides.More than 30% reduction in species richness was observed in the quadrats belonging to highest infestation grades when compared with that of the species richness of the no/negligible ones.Of the two components of species diversity, richness and evenness (Magurran 1988), the focus of invasive species impacts has been primarily on richness (Levine et al. 2003).Hulme and Bremner (2006) also reported that the presence of the invasive riparian plant Impatiens glandulifera resulted in more than a 25% reduction in alpha diversity.Similar decrease in species richness, diversity and evenness, at higher cover of some invasive species were also reported by Hejda et al. (2009).During this study, a significant negative impact of A. philoxeroides cover on associated native macrophyte richness was evident for all 3 seasons, even after adjusting for the influence of other invasive species.The present study also found a highly significant reduction in total number of associated native plants between the lowest (no/negligible) and highest infestation grades of A. philoxeroides.
We found negative impacts of A. philoxeroides cover on the total associated macrophyte diversity in littoral zones of aquatic ecosystems, though due to the lack of controlled/experimental study plots, it was not possible for us to conclude if the diversity of associated macrophyte species were directly affected by higher cover of A. philoxeroides or if it was due to some other factors.Some of the characteristics of invasive plants which are responsible for their strong impact on native plant communities are vigorous growth, dense rhizomatous structures, homogenous strands and high cover (Hejda et al. 2009), all of which are satisfied by A. philoxeroides.Aquatic A. philoxeroides has rapid growth and reproduction (Chatterjee and Dewanji 2012), high biomass accumulation potential (Zuo et al. 2012) and forms dense mats of entangled stems (Dugdale et al. 2010), some attributes which might help A. philoxeroides to pose serious threat to native plant diversity (Julien et al. 1995) by limiting the growth of other associated species in its close vicinity.Zuo et al. (2012) found a lower number of companion plants alongside A. philoxeroides under aquatic conditions when compared with its terrestrial ecotype, while Jin Cheng and Qiang (2006) also reported a gradual decrease in the species composition and diversity of the community with increasing dominance of A. philoxeroides.Bassett et al. (2012) also reported a strong competitive interaction between native vegetation and A. philoxeroides in the littoral region of a New Zealand lake, while acknowledging certain similar limitations of a field-based study.Further replicated research comparing uninvaded aquatic ecosystems with invaded ones; or greenhouse and/or field based experimental studies manipulating competition, would be essential for exact elucidation of the extent of A. philoxeroides impact on different plant communities.
The most commonly occurring macrophyte functional group was the 'free-floating' one, consisting of Eichhornia sp., Lemna sp., Spirodela sp and Pistia sp.; large growths of which are also known to adversely affect freshwater ecosystems (Scheffer et al. 2003).The consistent presence of other invasive macrophytes like Eichhornia sp., Pistia sp. and Mikania sp.across the different infestation grades of A. philoxeroides found in this study provides an ecologically important indication about the increasing presence of multiple invaders in aquatic ecosystems and suggest avenues  of future research on facilitative interactions between invasive species.Ricciardi (2001) reported that some invaders can alter habitat conditions in favour of other invaders, creating a positive feedback system and leading to accumulation of other non-indigenous/introduced species.Positive facilitative interactions between A. philoxeroides and other invasives like Eichhornia crassipes and Pistia stratiotes have also been recently reported (Wundrow et al. 2012).Aquatic and wetland plants are important sources of food, fodder, biomass and building material for human societies (Bornette and Puijalon 2011).Since people in South-East Asia spend as high as 72% of their income on food alone (FAO 2011), importance of the aquatic plants as cheap and quality sources of vegetables in Asian diet is huge (Peter 2013).Wetland plants as sources of traditional medicine provide health services to about 80% of the world's population, which further highlights their socioeconomical importance (Panda and Misra 2011).An additional impact of A. philoxeroides, which has not been widely demonstrated previously, is that A. philoxeroides reduces the availability of many plants useful for humans.Thus, as highlighted in this study, replacement of native aquatic/wetland plants, which are sources of cheap and popular food and medicinal sources, due to high infestation of A. philoxeroides, a less-popular food source in itself, is a huge cause of concern in terms of socio-economic sustainability of highly populated, developing countries of the world (Juffe-Bignoli et al. 2012).
Reports on quantitative assessment of impacts of A. philoxeroides invasion on native vegetation are mostly from temperate countries where the plant is naturally controlled during the winter season due to snow (Liu and Yu 2009).It is surprising that such studies from sub-tropical and tropical areas are severely lacking, considering its perennial growth in those regions and its status as one of the worst weeds of the world (Chatterjee and Dewanji 2012).However, accepting the limitations of an 'impact-only' field-based observational study conducted post-invasion, this study provides an insight into the floral associations of A. philoxeroides in this region and together with its potential to displace a large number of native flora at high infestation levels, it also highlights the socio-economic value of these plants, which is an unique feature to tropical regions.

Figure 2 .
Figure 2. The species accumulation curve for 192 quadrats for monsoon season.

Figure 3 .
Figure 3. Distribution of the cover percentage (%) of Alternanthera philoxeroides for each of the three sampling seasons (N= 192 quadrats/season).

Figure 4 .
Figure 4. Overall frequency of occurrence (%) of native associated macrophytes across the lowest (Grade -I) and highest (Grade -IV) of Alternanthera philoxeroides infestation.
Cook (1996))tification & common name sources:Prain (1903),Cook (1996), Ghosh (2005) * Plant general habit/provenance/status sources: http://plants.usda.gov/;http://www.issg.org/;http://www.iucnredlist.org/;http://www.season-effect(p-value=0.4484) on them, justifying the season-wise pooling of the quadrat data to estimate the effect of infestation of A. philoxeroides on the number of associated macrophyte species.Table2presents the associated macrophyte species seasonal presence/absence check-list, across the four infestation grades of A. philoxeroides.A general decreasing trend in number of associated macrophyte species was evident from grade I to grade IV of A. philoxeroides infestation for each sampling season.Significant differences across the four infestation grades of A. philoxeroides were also revealed by the Kruskal-Wallis H tests:

Table 2 .
Seasonal presence of associated macrophyte species across the four 'Infestation Grades' of Alternanthera philoxeroides for three sampling seasons.

Table 4 .
Socio-economic usage and their preliminary subjective valuation depicting the importance of some native plants and Alternanthera philoxeroides.